This is the second article in a series I’ve called “7 Fallacies About the Lengths of Things.” I’m answering seven questions in the diagram below, choosing from the possible answers on the right-hand side. In this article, I’ll consider the risk that government debt rises so high that an eventual default or restructuring becomes a certainty. Later, I’ll evaluate the amount of time that we have to fix our debt problem before we reach that point. In other words: “How Long Until it’s Too Late?” In my opinion, this is the most important question that we face in today’s economy.
How long until it’s too late?
Before a catastrophe strikes, there’s usually a time when you know it’s going to happen and that nothing can be done to stop it. You may be able to alter its effects but you can’t stop it. Think of a high speed train that’s gone off the tracks or an isolated boat that’s taking on water. And then apply the same idea to government debt. Once debt reaches a certain point, you know it won’t be fully paid off.
But how do you determine that point?
The approach that I’ll use here is to start with a hypothesis and then test it against the available history. History won’t settle the issue, but in this case it offers lessons that demand our attention. Here’s the initial hypothesis: Once debt exceeds 150% of annual economic output (GDP), it becomes certain that the debt won’t be fully paid off. Why 150%? I have three reasons:
- It’s a nice, round figure. For a task like this, there’s no point in being over-precise.
- It’s meaningfully higher than 90%, which is a key threshold identified in research by two of our foremost experts on government debt: Carmen Reinhart and Kenneth Rogoff. If 90% debt-to-GDP ratios implied future defaults, then Reinhart and Rogoff would have told us that. But they didn’t. They showed that us that economic performance usually worsens after debt exceeds 90% of GDP, without saying anything about the odds of default. (See here and here.)
- Conceptually, it seems that debt should become explosive when it reaches 150% of annual output. Not only does fiscal policy need to be extraordinarily tight to overcome the debt service burden while reducing the debt-to-GDP ratio, but other actions including a low interest rate policy are needed to keep funding costs down. It’s very difficult to maintain this balancing act for any length of time and through the ups and downs of the business cycle.
Testing the hypothesis
Using mostly Reinhart and Rogoff’s database, I count 51 instances of debt crossing above 150% of GDP in the last few hundred years. But some of these were part of the same struggle with high debt, or they lacked a full data set, or they lacked quality data and relevance because they happened in 18th or early 19th centuries. By screening out each of these categories, I’ll narrow the analysis to 26 cases of debt-to-GDP ratios breaching 150%.
Here they are:
My next step is to investigate what happened in each of these instances. Did the country default and, if so, how and when? Or, did it manage to reduce debt without default? What else was going on at the time, as far as inflation, politics, war and so on? Once again, I’ll start with Reinhart and Rogoff’s database and add information as needed.
My first conclusion is that the hypothesis is false, at least on a first assessment. Several countries successfully reduced their debt without defaulting or otherwise adjusting the amounts owed their creditors.
Notably, three of these instances occurred immediately after World War II. The United Kingdom’s debt reached 238% of GDP in 1947, while debt-to-GDP ratios for the Netherlands and Australia peaked in 1946 at 223% and 190%, respectively.
Does anything sound familiar about the timing of these peaks? Regular readers know that I devoted the first article in this series to debt reduction in America after World War II. America’s debt-to-GDP ratio didn’t exceed the 150% threshold, but it peaked at 122% in 1946 and then fell for the next three decades. And the United Kingdom, Netherlands and Australia lowered their debt ratios using roughly the same blueprint.
Based on their experience, here’s the proven strategy for winding down a debt-to-GDP ratio of over 150%:
Step 1 – Revert to peacetime levels of defense spending. Eliminate the deficit by slashing defense spending to near or below the “non-defense budget balance” (which is simply the budget balance excluding defense spending as defined in my last article and shown below for the United Kingdom). This is the most important piece. Without aggressively cutting the deficit, you won’t make much of a dent in your debt ratios. And if you continue to run large deficits the ratios will worsen. Fortunately, it wasn’t difficult for most nations to achieve a budget surplus after World War II. Soldiers returned to civilian life and factories were released from military use. European nations were further assisted by the Marshall Plan. In the five years beginning in 1948, the United Kingdom, Netherlands and Australia ran fiscal surpluses of 5.5%, 9.5% and
1.5% 3.0% (correction) of GDP, respectively. By 1953, only the United Kingdom’s debt remained greater than its GDP. The British then continued to run surpluses for the next twenty years by taking advantage of a large non-defense budget balance, as shown in the chart below.
Step 2 – Roll out the financial repression. If you continue to service debt at interest rates determined purely by market forces, borrowing costs may rise to unsustainable levels. To keep costs down, choose from a bag of tricks that could include capping interest rates, controlling capital flows to keep investors captive to domestic markets, requiring certain classes of investors to hold government bonds, penalizing other types of investments, and so on. These and other forms of financial repression were common in the post-World War II period. They helped to explain falling debt levels in the United Kingdom, Netherlands and Australia.
Step 3 – Inflate. Once you’ve controlled borrowing costs through financial repression, it’s safe to inflate away some of your debt. You’ll diminish the purchasing power of private wealth, slamming retirees and anyone else who lives on a fixed income. But your debt-to-GDP ratio will fall. Australia relied heavily on inflation to reduce its debt burden, with CPI data showing that prices doubled between 1946 and 1953. Inflation wasn’t quite as high in the United Kingdom and Netherlands, but contributed to declining debt-to-GDP ratios especially in the late 1940s and early 1950s. The annual inflation rate peaked at 15.9% in the United Kingdom and 13.6% in the Netherlands, with both peaks occurring in 1951.
Step 4 – Grow. Researchers Carmen Reinhart and M. Belen Sbrancia have argued persuasively that economic growth didn’t contribute to post-World War II debt reduction as much as the combination of financial repression and inflation. But still, it certainly helps to grow.
Understanding the blueprint that the United Kingdom, Netherlands and Australia followed, I’ll argue that it’s out of reach for America today. The argument is the same one I made in my last article. Essentially, the debt burdens of the late 1940s were caused by a burst in defense spending that was reversed after the war. By comparison, our debt problem today is explained by a growing gap between spending commitments that politicians won’t dare attempt to control and tax revenues that politicians won’t dare attempt to increase.
Therefore, the first and most important step toward debt reduction – restoring a budget balance – can’t be achieved this time by a surrender, truce and treaty.
Identifying the most relevant high debt episodes
Getting back to our list of high debt episodes from the table above, let’s set aside the three that were caused by World War II and look at what’s left. Three more of the remaining 23 haven’t ended in default, but once again there are qualifiers.
For both Hungary and Sri Lanka, default was averted by extensive support from the International Monetary Fund (IMF). We can surely dismiss these cases for lack of relevance to large countries.
That leaves Japan. Japan’s central government debt rose above 150% of GDP in 2004 and continued to increase, reaching 210% in 2012. Japan’s creditors may still rely on their government bonds, but in the same way that a child remains attached to the balloon that slipped out of her hands and now floats above the nearby trees. The balloon isn’t coming back, and holders of JGBs shouldn’t expect to be paid off in full. But I won’t complicate the analysis with my prediction; I’ll just remove Japan from the list on the grounds that an outcome hasn’t yet been determined.
Now for the conclusion: In every one of the remaining 20 episodes of debt rising above 150% of GDP, the original terms of that debt weren’t honored. Investors either stopped receiving interest in an outright default or they were coerced into a debt conversion. The years of default or conversion are shown in the tables below, which separate the particularly small countries (those with a global GDP share of less than 0.1% in 2010) from the rest. For the small countries, the debt threshold is almost always breached after the beginning of the default or conversion process, with the only exception being two cases in which the default and threshold breach occurred in the same year. In most of the larger countries, debt reached the 150% threshold before the default or conversion.
Three lessons of history
There are three key lessons that we can draw from this data. First, if America’s debt reaches 150% of GDP, there are no comparable episodes showing how such a large debt can ever be fully paid off. As I argued above, it seems that debt should become explosive at that point and there’s nothing in the data to suggest otherwise. Unless the excessive debt is explained by massive wartime spending that disappears after the war, or supported by IMF loans and assistance, it always leads to defaults or conversions.
And the empirical case is even stronger than suggested by these results, considering I’ve ignored the numerous instances of defaults occurring without debt reaching 150% of GDP. Reinhart and Rogoff’s data includes 197 default and conversion episodes in the last 150 years for the 69 countries they’ve researched. In the big picture including all of these events, we should recognize that:
- It’s unusual for countries to reach the 150% threshold before defaulting or converting.
- It takes a world war or the IMF for those few countries that do reach 150% to fix the problem without a default or conversion.
Together, these observations suggest that the chances of honoring your debt become virtually hopeless at above the 150% threshold.
The second lesson is not to believe the stories you might hear from conventional economists, who often dismiss the risks posed by our rising debt. Many economists would like you to think that a debt problem is like a trip to the candy store. You get to choose from a bunch of treats and no one forces you into anything. Let’s see… there’s the inflation option, the restructuring option, the tax hike option, the money printing option. With all these choices, you should just sit back and let your leaders make a wise decision. Think again. The reality is that once debt becomes explosive, default or conversion is inevitable and your government won’t be able to stop it, only to alter the way it plays out.
Which leads to the third lesson: Don’t look at inflation as an alternative to default, at least for severe debt burdens of greater than 150% of GDP. In these cases, inflation usually plays the role of a side dish. It tends to appear alongside a default or conversion and contributes to economic hardships but doesn’t prevent the default or conversion. Or, it appears alongside a budget surplus and helps to resolve the problem, as in Australia in the post-World War II period. But the key conditions for the problem’s resolution need to be present before inflation is added to the plate.
Setting aside the post-World War II period once again, inflation hasn’t prevented a default or conversion in any of the remaining high debt episodes. Several hyperinflations, such as 1920s Germany, stand outside the analysis due to incomplete data. But in these cases domestic debt was rendered worthless by the hyperinflation and corresponding currency collapse, while any external debt denominated in hard currencies was either defaulted on or restructured. And the defunct currency was eventually supplanted by a new one in a monetary system reset. Hyperinflation will certainly occur again in a world awash in debt. It’s easily created by central bankers who choose to monetize debt and remove any incentives for governments to act responsibly. But it’s equivalent to default.
Does the U.S. get a pass?
Now at this point in my article, readers with an “all is well” mentality may have already dismissed the relevance of these results. They might argue that America will make good on its debt because “it always does.” Or they’ll point confidently to America’s unique advantages as a military superpower, paragon of political stability and steward of the world’s predominant reserve currency. Confronted with the lessons of history, they’ll say “this time is different.” But what exactly is it that may or may not be different? It’s important to draw a distinction between two concepts of debt limits:
- The point-of-no-return. At what point does it become virtually certain that a debt problem won’t be resolved without an eventual default or conversion?
- The overall debt tolerance. At what point does the ability to bear debt completely break down and actually trigger the default, conversion or hyperinflationary money printing?
To answer the second question first, the U.S. certainly has a higher debt tolerance than most and possibly all of the rest of the world. Contingent on what else is happening in the world, our debt could probably blast through 150% of GDP without people taking much notice. We’d see a headline or two and go on with our lives.
But the research that I presented here speaks to the first definition, not the second. It gives us an estimate for the point at which the train goes off the tracks, after which it can roll on for quite a while before it hits the tree line. But it won’t just roll right back onto the tracks and continue on its way. Unfortunately, the existence of this critical threshold is rarely acknowledged and almost never discussed. Which is one of the major reasons that countries so frequently sail through it unknowingly.
Another important distinction is that the point-of-no-return shouldn’t be as country-specific as the overall debt tolerance. It’s based on the mostly mathematical but partly political impediments to lowering debt once it’s risen beyond a certain amount.
You can think of it as a Catch-22. Once you reach the point-of-no-return, then the only way to get back to safe debt levels is through austerity. But austerity can have the opposite effect because it throws the economy into reverse and fiscal measures that are normally deficit-reducing can turn out to be deficit-increasing. Not to mention the fact that austerity kills political careers. The key is to stay clear of the point-of-no-return in the first place, because once you’ve reached it you’re toast.
I propose that there’s no such thing as immunity from this Catch-22. My reasoning is that the mathematics and politics of austerity aren’t that much different from one country to the next. Therefore, America’s point-of-no-return isn’t that much different than it is for other large, developed nations. We’ll never know for sure, because we’re talking about a hypothetical level that can’t be measured precisely. You may even note the shenanigans in Washington and conclude that we’re already past it. But for the purposes of this article, I’ll be optimistic for a change and accept my initial hypothesis – 150% of GDP is the point-of-no-return – which also happens to be consistent with my interpretation of the historical results.
To be continued…
Getting back to the length fallacy challenge, my conclusions only take me half way to an answer to the question: “How Long Until It’s Too Late?” In a later article in this series, I’ll complete the analysis and choose from the list of possible time frames. Between now and then, I’ll publish additional articles that’ll help me with that choice. And for now, I’ll record my intermediate answer: “Once our debt reaches 150% of GDP.”
Data sources and reading recommendations
Research on the relationship between debt and growth
- Carmen M. Reinhart and Kenneth S. Rogoff, “Growth in a Time of Debt,” American Economic Review 100 (2010): 573-78.
- Carmen M. Reinhart, Vincent R. Reinhart and Kenneth S. Rogoff, “Debt Overhangs: Past and Present,” Journal of Economic Perspectives (forthcoming).
Fiscal and inflation data
- British budget data is sourced from: B. R. Mitchell, British Historical Statistics (Cambridge: Cambridge University Press, 1988).
- Australian budget data is sourced from: B.R. Mitchell, International Historical Statistics: Africa, Asia & Oceania 1750-1993 (London: Macmillan, 1988).
- For the Dutch fiscal balances, Frits Bos kindly supplied me the data that he used in: Frits Bos, “The Dutch Fiscal Framework: History, Current Practice and the Role of the Central Planning Bureau,” OECD Journal on Budgeting Vol. 8 No. 1, 2008.
- For inflation, I used data that Carmen M. Reinhart and Kenneth S. Rogoff have generously made available from their bestseller: Carmen M. Reinhart and Kenneth S. Rogoff, This Time is Different: Eight Centuries of Financial Folly (Princeton: Princeton University Press, 2009).
Research on the role of repression and inflation in post-World War II debt reduction
- Carmen M. Reinhart and M. Belen Sbrancia, “The Liquidation of Government Debt,” NBER Working Paper No. 16893, March 2011.
Debt conversions summarized in Table 2
- 1881 bond restructuring in Turkey: Christopher Clay, Gold for the Sultan: Western Bankers and Ottoman Finance, 1856-1881 (London: I.B. Taurus, 2000).
- Late 19th century debt restructurings in Spain and Greece: Marc Flandreau, Frederic Zumer, The Making of Global Finance 1880 – 1913, Development Center of the OECD, 2004.
- Chilean debt restructurings in 1936 and 1948: Erika Jorgensen and Jeffrey Sachs, “Default and Renegotiation of Latin American Foreign Bonds in the Interwar Period,” NBER Working Paper No. 2636, June 1988.
- 1926 Italian debt conversion: Silvana Bartoletto, Bruno Chiarini and Elizabetta Marzano, “The Sustainability of Fiscal Policy in Italy: A Long-term Perspective,” CESifo Working Paper No. 3812, May 2012.
- 1930s debt conversions in France, Italy, New Zealand and the U.K.: Carmen M. Reinhart and M. Belen Sbrancia, “The Liquidation of Government Debt,” NBER Working Paper No. 16893, March 2011.
- 1990s debt restructurings in Algeria: Karim Nashashibi, Patricia Alonso-Gamo, Stephania Bazzoni, Alain Feler, Nicole Laframboise and Sebastian Paris Horvitz, Algeria: Stabilization and Transition to Market, IMF Occasional Paper 165, 1998.
- Haircut estimates on 2005 Argentine and 2012 Greek bond swaps: Jeromin Zettlemeyer, Christof Trebesch, G. Mitu Gulati, “The Greek Debt Exchange: An Autopsy,” draft dated September 11, 2012.
 For example, a country may convert its existing bonds to new bonds and allow the debt ratio to return to the 150% threshold but with the new bonds paying a below-market interest rate. In other words, the country is no longer technically in default but it’s still enjoying the benefits of the prior conversion. In this case, I’ve ignored the fluctuations around the 150% threshold and called it a single struggle with high debt.
 Debt/GDP estimates for the 18th and early 19th centuries are available for a few European nations, but these nations controlled foreign colonies with substantial resources that were often plundered and wealth that was heavily taxed. And they acted as bankers to these colonies, raising funds domestically to invest in their domination, administration and exploitation. (I know, too many -tions.) In other words, much of the public debt in, say, England and the Netherlands, was easily serviced through imperialism. And it wasn’t constrained by the size of the domestic economy. To develop metrics comparable to the debt/GDP ratios reported today, you would need to account for economic activity throughout an empire, and I’m unaware of any research that does this effectively.
 An increasing debt-to-GDP ratio after a default can be explained by a variety of factors, such as: declining GDP; currency devaluations that push up the value of foreign currency-denominated debt; more honest accounting practices once governments stop trying to convince creditors that they won’t default; additional funding from creditors seeking to protect their interests by helping to avert further economic and political turmoil (or creditors simply doing what the government tells them to do); and bailout loans from the IMF or other sources.